,

$PE

6721

Predicting Phase Behavior of

Condensate/Crude-Oil Systems Using Methane
Interaction Coefficients
D. L. Katz, SP&AIME, U. of Michigan
A. Fuoozabadi, Abadan Instituteof Technology

Introduction
Prediction of phase behavior of complex hydrocarbon
systems has improved gradually. Recent papets using
equations of state for predicting phase fugacities of constituents have extended their usefulnessto higher boiling
compmenfs .1-5Using extended analysis with boiling
points of higher boiling hydrocarbon groups and their
densities and molecular weights provides a basis for
improvement when predicting phase behavior. At the
high pressureof reservoir condkions, properties of these
Cd+groupsare impor~t.
Using the Peng-Robutson5equation of state has the
advantageof improving accuracywhencalculatingliquid
density. When used by Bergman et af.46with extended
gas analyses, the equation gave good prediction of dew
points and liquid yield at pipelineconditions. Bergmanet
al, used interaction coefficients but did not completely
explore their usage since phase behavior of condensates
at pipeline conditions appeared rather insensitive to the
Co+interactioncoefficients.
This paper reports the extension of the PengRobinson-AGA procedure for phase prediction of condensates and natuml-gas/crude-oil systems at reservoir
wnditions. New interaction coefficients were found
necessary for methane and Ce+ constituents (especially
the highest boiling group, such as Cm+)so that experimental behavior could be matched by prediction. Steps
for developing interaction parameters are presented.
Then, agreementbetween measuredand predicted phase
0148-21%/78/001 1-6721$0025
@197880ciety o! Pemleum Enghews

OfAME

behavior up to pressures of 64121 kpa (9,300 psia) is

shown for several crude oil and condensate systems.

Calculations of Phase Behavior

The equationof state, using the relationshipof fugacity to
partial molal volumes, is used to compute the fugacity of
each component !n both liquid and gaseous phases. At
equilibrium the fugacity of each component in liquid
equals its fugacity in the gaseous state.
Basic parameters for each constituent are the critical
properties, normal boiling point or vapor pressure. For
hextme and heavier groups present in natural-gas and
crudeoil systems, average normal boiling point, liquid
density, and molecular weight are used7instead of measured properties available for pure components.
For systemscontaining variousmolecularfamilies, the
interactioncoefficients between the primary constituents
must be includedto accountfor the degree of compatibility. Here methane and higher boiling C6+groups are the
principal components for which interaction coefficients
appear sigtilcant.
Since propertiesof Cd+groupsin crudeoil and ccmdensates are needed here, generalizedproperties are given in
the absence of measured values.

Properties of Hexanes and Heavier Groups

Average boiling @nts, liquid densities, and molecular
weights were reported by Bergman et al, 8 from C6
through C15,based on analysis of 26 condensates and
naturallyrecurring liquid hydrocarbons. This correlation
I

Thispaper extends the Peng-Robinson-AGA procedure forcomputingphase behavior to include

gas-coruknsate and crude-oil systems at reservoir conditions. Irnwaction coeficientsfiom
wethane to the C6+ groups were correlated with liquid density. With these coefficients, the
behavior of gas condimsate systems maybe predicted reliably.
NOVEMBER, 1978

1649

. .
500

1/ I

0Wgmon,Tek EkIGatz
xLiquid

0.s

as

[,//,
-50

1,/,

100

J=-.<. A====iJ

..-.

200

,11

1,!

300

400

500

v.mr~ffm,nc~ump~H.c~t+/;-;

I
-100

was extended to a normal boiling point of 5WC using

added data8e for use with reservoir condensates and
crudeoils. Table 1and Figs, 1and 2 give average properties for groups of compounds boiling between 0.5C
above the previous n-paraffinsl and O.5C above the
carbon number used to identify the group, The higher
boiling point properties are only best estimates and are
used for systems for which propefiies and/or analyses
have not been performed at these high boiling points,

100

300

F@.

2
~
C,*
G3
C*4
::

c,?
:,:
C.zll
C2,
G42
$4
C*,
c27
C*B
$
Gw
24
CM
C3e
C37
$8

c,,
C42
c.
C.sl
C4S
1650

Soiling Range
Vc)
36.5 to 69.2
69.2 to 96.0
$8.0 to 126,1
126.1to 151,3
1S1.3 to 174,6
174.6to 196.4
186,4to 216.6
216,6 to 235,9
235.0to 253,9
253.9 to 271.1
271,1 to 287.3
287 to 303
303 to 317
317 to331
331 to 344
344 to 357
357 to 369
369 to 361
381 to 392
302 to 402
402 to 413
413 to 423
423 to 432
432 to 441
441 to 450
450 to 459
459 to 488
466 to 476
476 to 483
463 to 491

w)
97.9 to 156.7
156.7to 210.1
210,1to 259.1
259.1 to 304.4
304.4 to346.4
346.4to 365.5
365.5 to 422.2
422.2 to 456.7
456+7to 469,2
469.2 to 520
520 to 547
547 to 577
577 to 603
603 to 628
62~ to 652
652 to 675
675 to 698
696 to 717
7f7 to 737
737 to 758
756 to 775
775 to 703
793 to 810
810 to 626
626 to 842
642 to 657
657 to 874
674 to 686
668 to WI
901 to 915

!
400

I
5C9

2Molecu!&rw boiling pointsof hydrocarbongroups in

crude oil and condensatesystems,

Development of Interaction Coefficients

Initial trials of predicting bubble points of crude oils with
the Peng-Robinson-AGAprocedure failed to give good
agreement, even though extended analyses up to C20+
were used. The interaction coefficient, used earlier by
Zudkevitch and Joffe,ll was inserted in the equation of
state by Peng-Robinson.s Using alternative interaction
coefficients showed promise as a way of modifying the

TABLE 1GENERALIZED PROPERTIESOF PETROLEUMHEXANEPLUS GROUPS

Hydrooertron
Group
c,
c,

NORMAL%?LING POINT, W

NORMAL BOILING POINT, %

Fig. lDensity vs boiling Wlnts of hydrocarbongroups Incrude

oil and condensatesystems,

!__i

-4.-

100

Average
Boiling Point
~c)
~F)
63.9
147
91.9 197,5
116.7
242
142.2
266
165.8
330.5
187.2
369
407
208.3
441
227.2
475.5
246.4
266
511
263
542
572
%
505
617
325
336
640.5
351
363
E
707
375
727
366
747
397
766
419
429
::
817
438
455
463
471
478
466
493
500
506
515
522
528
540
rd

r:
r5
808
022
::
050
972
082
993
1,004
1,017
1,027

DensHy
(g/ml)
0.665
0.722
0.745
0.784
0.778
0.769
0.800
0.611
0.622
0.632
0.839
0.847
0.852
0.857
0,662
0,667
0.672
0.877
0.881
0.665
0s69
0.893
0.896
0.838
0.902
0.906
0.309
0.912
0,014
0.917
0.919
0.922
0.924
0.026
0.028
0.930
0.931
0.933
0.935
0.937

Moleoular
Weight
64
96

107
121
134
147
181
175
190
206
222
237
251
263
275
291
R
E5
::
402
416
430
456
472
466
500
514
528
542
656
570
864
598
612
628

procedure. Accordingly, interaction coefficients for

methane binaries were compiled based on bubble points
along withdescriptivepropertiesthat might allow a correlation related to known propertiesof the C8+groups. The
interactioncoefficients obtained are the values fhat gave
minimum error for the bubble point for the range of
temperatm and pressure indicated in Table 2, using data
from Refs. 12through 20.
The use of liquid density at 15.5C (60F) for the
binary component in the methane binary gave a correlation of the interaction coefficients for these systems
(Fig. 3).
Interactioncoefficientsfornonhydrocarbom used here
are 0.12 for nitrogenhydrocarbons,2021
and 0.1 SforC02
hydrocarbons.8 Ethane and propane intel action
coefficient were considered with other hydrocmbons.
Because of the variation found with temperatun! and
pressure, and the relatively small influence on the predictions, the value of 0,01 used in the Peng-Robinson-AGA
procedure was continued. The interaction coefficients
between nitihane and ethane through n-pentane were
zero. Later, the significance of the interaction coefflcients for methane and the last composite group, such as
Cm+, will be discussed.

Natural-Gas Condensate Systems

natural-gascondensatesystem used whencomparing
experimentaldata with calculated valuesis listed in Table
3. The measured dew-point pressure at stated temperatures, the liquid yield when flashed at given temperatures
and pressures, and the corresponding equilibrium ratios
are used to give three methods for comparison. For each
system, the extended analysis was available. The predictions for the condensates were based on the interaction
coefficients of Fig. 3, which is correlated on liquid density of the hydrocarbon groups as developed here. For
natural-gas condensates, none of the experimental data
being compared were used to generate the calculation
procedure. Each comparisonwill be discussed in turn.

The

005

I
0.60

070
OENSITY AT 15.5~ghnf

I
a90

Lo

fl . 3lnteractbn coefficient from methaneto other

hyr?rocarbonsas a functbn of the densityof the iwavler
hydrocarbon,

Hoffman et al. 8Lean-Gas Condensate

Hoffman et al.s data on the gas-cap material at the
reservoir temperature of 93, 9C (201F) are the most
complete because true boiling-point distillations with
densities and molecularweights were determinedon cute
corresponding to n-paraffins up to C22. The analysis
given and the densities and molecular weights for each
cut were used to predict(1) dew-point pressure,(2) liquid
yield as a function of pressure down to 3551 kFa (515
psia), and (3) vapor-liquid equilibrium ratios all at
93,90C, Table 4 compares experimental data, predicted
values in this study, and, published predictions by
Starling.i Fig. 4 compares equilibrium ratios. The dewpoint pressure and liquid-yiekiprpdictions are excellent,
end predicted equilibrium ratibs for methane have an
average absolute error of 0.76% with experimentaldata.

Starlingl Rkh-Gas Condensate (Fluid B)

Starling reported the analysis on a rich-gas condensate

TABLE 2INTERACTIONCOEFFICIENTSFOR BINARY METHANE/HYDROCAftBON SYSTEMS

Heavier
Comwnent
nC4
nC8
nCa
G
nc,
rlc*
nClo
a
E??nzene
Cycbhexerw

Intereotbn
Coeffwnt
0.02
0.02
0.025
0.025
0.036
0.035
0.035
0.054
0.08
0.03

Boning
Point
(c)
-0.50
38.07
68,74
88.43
126.86
150.80
174.12
343,8
80.03
80.72

D&y
1S,6C)

0.584
0.6312
0.6F40
0.5882
0.7068
0.7217
0.7341
0.795
0.6s44
0.7834

Moiecuiar
Weight
58.12
72.15
86.18
100.20
114.23
128.26
142.28
282.57
78.11
84.16

Pressure
Range
(kPa)
4100to12400
4100to16100
4050 to 16200
4100t020800
4050 to 7080
4050 to 30400
4100 to 32750
500 to 8050
9970 to 24200
4100t024 100

Temperature
Range
(=c)
71.1,87.8,104.4,121.1
71.1,104,4,137.8
75,100,150
71.1,104.4,137.8
75,100,150
75,100,160
71.1,104.4,137.8
40
37.8,71.1,104.4
71,1,104.4,137.8

Ref.
12
12
13
14
15
16
12
17
18,19
20

TABW=~wwNm~/CRUM~LSwTUSUwDIN
COMPARlSt3NSOFEXPERlMENTALWl~
PREDfCTEDBEHAVtOR
Authors
Hoffman,etal.s
Starting!
Rohrl&J
stalKwJf and
Katz end
l+aohmut~
Evarmand Harriet
NOVEMBER, 1978

System

Temperature Range
~c)
~F)

Pressure Range
(kPa x 10+)
(paia)

Condensate,crude
83.8B
10JIJ glJ
.,.

201
214,250
120,200

26.43 to 3.55
26.16 to 3.55
7.22 to 64.63

3,842to 616
3,705 to 515
1,047to 9,374

crude Ofl

48.88,121.11

120,250

6.26 to 38.13

1,010to 6,530

Crude oil
Crudeo!i

03.33
87.78

200
1s0

0.813to 16.54
7.65 to 41.37

118to2,333
l,l10t06,000

C&%eate
crude on

16S1

TABLE4-COMPARISON

OF PREDICTEt)WtTH EXPERIMENTAL DATA FOR GAS-OAPCDMP~lllm

HOFFMAN ot8L@AT 36.9%

OF

Ylsld*
Starlingl
This work
(rnVmaxlOo) (trbl/MMsef) (n$/mxlO$)
(bbl/MMaef)
64.60
11.3
Q.60
63.66
70.26
12.49
77.62
13.8
84.40
15.02
90.66
1601
92.87
16,51
96.44
17.6
04,50
16.60
17.3
97.31
88.00
15.25
14.0
83.81
Experimentaldew-poht pressureis36430 kPa (3,642 psia),
Predicteddew pokrtforthis studyis26682 kPa (3,670 pale),Starllngt>35267 kPa (6,116 psia).
Averageabsolutedeviationbetweenexperimentaland pradictsdylslds forthisstudy 1s2.6%,and8,2% for Starling.
Ltquki

Pressure
(kP@
(pals)

20098
2,915
17340
2,516
13633
2,016
19446
1,515
8998
1,015
515
3551

Expsrtmental
(rrrVmxlOe) (bbVMMsof)
51,02
9.07
69.97
12.44
67,52
15.56
95.51
16,08
95.29
16,94
84.82
15.08

Llquldbulunwal raprtad pmawe andtempaa!urw oaa standardcrx@kmafor nwtrk unkaare 15Cand 101.326 kPa.

with extended analysis to Cm, but did not include liquid

densities and molecular weights for the groups. His
analysis with generalized liquid densities and moleeular
weights of Table 1 were used with the interaction coeftlcienta in Fig. 3. Dew-point measurements at temperatures of 101.1*C am 121.1C (214F and 25(PF) are
wmpaiwd with predictions along with liquid yields at
these temperatures in Table 5. The method in this study
showed no improvement when predicting dew-point
pressure, but a modest improvement when predicting
liquid yieId, Note that the Starling method involved rePRESSLJRE
, Psio

0(R& ~to)

gression analysis on measured equilibrium ratios to determine parameters for the common Bendict-WebbRubm equation of state and then predictions were made
on the same systems.

Crude-OiUNatural-GasSystems
Dataavailableforcrude-oil/natum!-gassystems are listed
in Table 3. Before making phase calculations, extended
analyseshad to be preparedfor some systems by converting distillations to individual hydrocarbon groups using the generalized properties in Table 1. Density and
molecular weight of the last fraction were adjusted
to provide an over-all match of reported density and
molecularweight of the ~+ fmction.
Pteliminaq trials when predicting the behavior of
crude-oilhatural-gas systems indicated that the interaction coefficient relationship with density did not give
satisfactoryresults whenextendedto the last fraction. We
knew that asphrdticand high-boiling constituents occurring in the last fntction, such as C20+,were not likely
to have a continuous volatility behavior when measured
in terms of liquid density. In the work of Rokmdzqand
Standing and Katz?$the downward curvature of the
vapor-liquid equilibrium ratios with increased pressure
was unexpected. At that time, the explanation was that
concentrationof the high-boiling and asphaltic fractions
in the liquid state by retrograde vaporizationof the intermediate constituents was responsible for the behavioq
equilibriumcell contents were approachingdew points at
increasedpressure. One might, therefore, expeet to have
a discontinuity in the interaction-coefficient/density
function as the last constituent is reached. Accordingly,
tineinteraction coefficient for the last group was determined from even a single reliable point (a bubble point,
for example)because its use for predicting equilibria for
other conditions is shown to be reliable.

kJ5L=c

1.0

Prsdktsd:
C3

-c{a
-%.+]

~
,01
E
m
-.

%==+++
[j

; /:

Hoffman et aLs Crude-Oil System

..,L-Ju&JJ

40,000
PRESSURE;kpo
Fig. 4Comparisonof prediotedwith experimentalK values of
Hoffman
etal,condensateat 93.WC(201F),
1ooo

1652

Data are composed of composition of the equilibrium

liquid and gas phases at93.!PC(201F)and26455 kPa
(3,837 psia)at reservoirconditions. The interactioncoefficient of the last group, Cm+, is computed using the
bubblepoint pressure of the liquid of 0.17. Using this
value, the gas-condensate composition calculatedas part
of the bubblqxint computation is compared with the
measuredcomposition of the gas-cap fluid in Table 6.

Roland23Natural-GatdCrude-OilSystem
Based on the distillation curve, density and molecular
JOURNALOF PETROLEUM
T2CHNOU)OY

TABLE 5-COMPARISON

Pressure
(kPa)
(peia)
G
2,015
10445
1,515
6928
1,015
3661
515
13893
2,016
8722
1,265

OF PREDtCTED WtTH EXPERIMENTAL DATA FOR STARLING CONDENSATE AT

TEMPERATURES OF 101.1and 121.lC (214mMf 28(FC)

Temperature
(%)
101.1
101.1
101.1
101.1
121.1
121.1

Experimental
(rn/m3xl&)
(bbl/MMsof)
166.5
880.2
156.0
677.5
145.3
822.9
115,3
848.8
119.5
672.2
126,4
711.0

Measured(kPa)
Oew-PointTemperature
(214F)
26156 (3,795 psla)
101,1C
(250F)
24235 (3,515 P$ia)
121.lC
Lk@dKIkmraal rermrtedmeawre andterrrpwature,and gas srandardcmritbmsfw

LiquidYield
Thie work

Starling

(m/mxl&)
(bbi/MMeof)
148.4
823.60
136.8
780.7
125.8
706.5
102.4
576.0
122.7
6S0.2
112.7
633.9
This Study (kPa)
26641
25788

(m/m$x109
1058.9
390.0
872.4
872.4
830.8
755.4

(bbl/MMsof)
187.9
178.0
165.1
119.7
147.7
134.3

Starling(kPa)
24 235to25614
23 548to24235

metricunts are 15C ard 101.326 kPa. Avers@ absolutedetia!bn kmthh studyla9.4/.

wit 12.1/0forStarting.

weight of the C7+residue, and the generalized properties

in Table 1, an extended analysis up to the ~u+ group was

conducted for the crude oil of constant composite composition of Roland.m
The interaction coefficient for methane through C=+
was computed by adjusting methane K values for the
24587-kPa (3,566 psia) and 38480-kPa (5,581 psia)
runs at 1200F.Tne equilibrium ratios of all other runs of
constant composition then were computed. FYedicted
equilibrium ratios at 93.3C (2WF) are compared with
experimentaldata in Flg. 5. The average absolute deviation of 14runs at 93.3C (200F)for methane K values is
2.30%. However, if the K value of the runat54917 kPa

PRESSURE, psio

10.0

111111 t 1I1
500 1000

200
i

c1

.D Experimen?ol
-

Colculoted

10,000

Ii

-,~.
1

1.

o J

C4

C5

-i
~
E

m
~
2w

~:

,1:0

0.1 ~

Stsnding and Katz24Natural-Gas/Crude-Oil

System
Six runs were selected fmm Standing and Katz data.
These runs belong to 49,9API crude oil with a GOR of
650 mVm3 (3,660 scf/bbl) at 48 .9C ( 120F) and
121.1C(250F). Standing28gives extended analysis for
crude oil with cut intervals of 37.8C (100F), and also
data for molecular weight and density of high-boiling
constituents vs boiling point.
The interaction coefficient between methane and the
last group, with boiling points starting at 371. lC
(7000F),was obtained by using the methane K values for
two runs at 121.1C (250F) w~thpressures of 21960
kPa (3,185 psia) and 29715 kPa (4,310 psia). Fig. 6
compares predicted with experimental K values for three
runs at 48.9C (120F).The proposed pweedure is capable of predicting downward curvature of ethane through
heptane plus K values at high pressures, Deviation of the
K values is given in Table 7.

b/
-7

______.;;

I .0

3000

(7,965 psia), which seems to be in some e~or~ is excluded, the absolute deviation will improve to 1.68%.
Furthermore, for 13 runs of 48.9C (120F)and 93.3C
(2WF) and below 46471 kPa (6,740 psia). absolute
deviation for methane equilibrium ratios is only 0.83%.
MethaneK values and componentsthrough Cc+(Table7)
indicate excellent agreement between predicted and
measured values.

1,

.7,

-C61<7

-!!

~.

.:

Katz and Hachmuth25

Crude-Oil/Natural-Gas System
Although the composite composition is not provided for
this set of early data on the equilitilum ratios of the

-.):0
,,4

,,

TABLES-COMPARISON
OF PREDICTED WITH MEASURED
GAS-CAP PHASE COMPOSlllON WITH HOFFMANot s/.-

\jl/

0.01

0.001.
moo

.--l
c;

.+

.;-;.?

I
10,000

I
!00,000

FIWSSLfRE . kPo

Fig. 5Comparbon of prediotedwith ex rimentaiK vaiuesfor

Rolandmcrde oil at 93.9 (?k+:~:
(20tTF).
NOVEMBER, 1978

Component
c,
0s
G
nC&
Go

C15

Measured
Composition
of Gas-Cap
Well Effiuent
(moi%)
91.35
4.03
1.5s
0.19
0.156
0.060

Bubble Point
Calculatbn on
Crude Oil
(MO!%)
92.12
3,504
1.52
0.16
0.116
0.046
0.0036
00019
1653

TABLE 7AVERAGE ASSOLUTE PERCENTAGE DEVIATtON FOR METHANE TNROUGH

C,+ K VALUES FOR C13UDE-OILSYSTEMS

Component
Roland=
Standingand Katz
c,
2.4*
2.2
c,
3.0
4.8
3.4
G
8.1
c,
&o
c,
4.3
1::
cc
8.7
13.6
c,+
1Q.8
18.9
Equilibria
compared
20
8
Average abaolutadetiatbns aa pefoemtofexpadnmntalva!uva

crude-oil/natural-gas system, the smoothness of the

measuredK values indicates that the compositecomposition change is negligible. Based on this assumption, one
can compute the composite composition in the cell.
Vapor- and liquid-phase compositions of the run at
12024 kPa (1,744 psia) and 93.3C (20(YF)were selectedto compute the composite composition. Liquid and
vapor phasesof this run were combined as 2 mol gas and
1mol liquid, Extended analysis was conducted the same
way as for Roland crude oil. The interaction coefficient
for methaneand the last fxaction(i.e., ~s+) wasobtained
by matching predicted vaporizatitxr to equal 66,67%,
With this value of interaction coefficient;,K values were
predicted for other rur at 93.3C (200F).Table7indicates that predicted values agree well with experimental
measurements.

PRESSURE,
psia

10.0
t

1,000

KJo

Katz and HachmuthZO

2.3
2.3
3.6
F3
10.1
33.5,

Evansand Harrk$
28
9.6
6.8
10,3
13.1
13.9
37.2

10

Evans and Harrisg

Crude-Oil/Natural-Gas System
Extended analysis for the crude oil was conducted for up
to CM+;the intemction coefficient between this last fmction and methane was obtained by adjusting methane K
values for runs of Mixture B. Fig. 7 compares predicted
with experimental K values for constituents of Mixture
A, and Table 7 shows percentagedeviations,

General Comparison
Table 8 assembles the interaction coefficient between
methane and the last group for crude oils, accompanied
by properties of the last fraction. All extended analysis,
except for Hoffman et af, crude oil, did not have true
boiling-point distillaticm analyses. Extended analyses
were improvised using judgment and generalized ex-

10.0
.

1,000

PRESSURE,
psio
2,000
1

10,000

II
Experimentoi
Predi \d

o 0

c,

c1

Lo
x
g
5 :7
a
~
~ 01
m

~
3
g

\
c, -b
1,,
c30

,.0

C4

CCJ

t -

i:
ce

..._ I

o.cq,&_wJL

o u Experiment!

Predicted

~y-.___+
T
--.

Io.000
[00.000
PRESSURE,kf%
Fig. 6CamparLsonof redictedwith expermentelK valuesfor
650-ma/m(3,680scf/b E1)gaa-to~il ratioof Standingand Katzz
crude oil at48,9W(120F).
1654

~6/

Hc~
I

--

0.01
-

c%di

0.1

.p

Cy

o.(),~

100,000
PRESSURE,
kPo
R . 7Ccmparlson of radictedwith experimentalK vaiueafor
$lixtureAcrudeoilo~EvansandHarrisO at87.6C(1900F).
JOURNALOF PETROLEUMTECHNOLOGY

COEFFtOtENTSBETWEENMETHANEAND MST
FRACTIONOPCRUDE=OILSYSIEM8

TABLE O-INTERACTION

Crude Oil
Hoflmanetal.s

Fr~@&r
G+

Rolan&

Go+

standngandKatz*
KatzandHaohmutffs
EvansandHarrbs

Css+
ox!+
%8+

Interactbn
Coefficient
0.17
0.17
0.25
0.20
0.25

tendedhexane plus its propetiies. Therefore, any correlation of methane with the last groups interaction coefficient should wait for more data with systems, including
true boiling-point analyses.

Recommendation
When obtaining composition data for crude-oil reservoirs, bubble-point presswe customarily is obtained at
the reservoir temperature on a bottom-hole or recombined sample. The composition with extended analyses
to GO+ or more, using a tme boiiing distillation along
with bubble point, Peng-Robinson-AGAprocedure, and
our proposed technique (including properties of the
groups, if necessary)can be used to assess the interaction
coefficientfor methane and the last fiaction.The interaction coefficient then will give the information required
for predicting phase behavior at other conditions.

Conclusions
Proper estimation of methane K values in both crude-oil
and condensate systems is essenthd for an accurate estimate of bubble point, dew point, and flashhtg of these
systems. UsirtSour proposedprocedure, predicted methane K values agree well with measured Vah-es.
Forgas-condensate systems, the Peng-R.,. i m-AGA
procedure with extended an~ysis and interact!..2 coefficients for methane and C6+constituents correlated with
density seems to give reliable results. For crudeoil/
natural-gas systems, the interaction coefficient for methane and the last fraction shottld be determined with
a bubble @rtt, or selected from Table 8. Gathering
data with true boiling-point analyses of crude oils should
help correlate the interaction coefficients for the last
fraction containing the high boiling-point material and
mphalt.
References
1. Starling, K. E.: *ANew Approachfor L%tmminingEquation-ofState ParametersUsing Phase EquilibriaDcta, Sac. Pet. Eng.J.
(Dw. 1966)363-371;Trans., AIME, 237.
2. Soave, G.: *EquilibriumConstants From a Modified RedlichKwong Equation of State, Citenr. Eng. Sci. (1972) 27,
1197-1203.
3, West, E. H. and Erbar, J. H.: *AnEvaluationof Four Methodsof
Predictingthe Thermodyrramicproperties of Light Hydmcasbon
System, Proc., NGPA 52nd AnnurdConvention]:,Dallas (1973)
56-61.
4. Bcrgrnan,D. F.: P,tilcting the PhaseBchaviorof NaturalGsrsin
pipelines, PhD dissertation, U. of Michigan,Ann Arbor (1976).
S. Pcrtg, D. Y. and Robinson, D. B.: ANew Two-ConstantEqution of State,*Ind. Eng. Chem. Fund. (1976)15, No. 1,59-64.
6. Bergman; D. F., Tek, M. R., and Katz, D. L.: Retrograde Condensation in Natural Gas Pipeiines, MonographSeries, American
G= Association, New York City (1%5).
7. Cavett, R. H.: Physicsl Data for DhNillauon Calculations
Vapor-LiquidE@Ubna, Prac., API 27th Mid-YearMeeting
(1%2)52,351-366.
NOVEMBER, 1978

Last Fraotlon
Molsoular
Density,at 15J5C
(g/ml)
m
0.202
361
0.028
,
492
1.008

0.930
%
0.914
342
.
8. Hoftinarm,A. E., Crump, J. S., and HocQtt,C. R.: Equilibrium
Constants for a Gas-CondensateSystem, Trans., AIME (1953)
198.1-10.
--.--9. Evans, R. B. and Harris, D.: 2quiJibrium VaporizationRatiosHydmcsubon Mixture Containing TWOConcentrations of Heptanes and Heavier Fkacrions, Ind. Eng. Chem. (19S6) 1, No. 1,
45-50.
10. Selected Values of Ftoperties of Hydrocarbons and Related
Compounds, API Reject 44, TexasA&M U,, CollegeStation
(1%9).
11. Zudkevitch, D. and Jaffe. J.: Correlation and Reoiction of
Vapor-fiquid Equilibria with the Redhch-Kwong Equation of
State,AfCitEJour. (1970) 16, N0. 1,112,
12. tigc, B, H. md Lacey, W. N.: Monograph on API Research
Project 37 Thermadjwatnic Properties of the Light Par@in
Hydrocarbons and Nitrogen, API, Dallas (1950).

13. Shim, J. cmdKohn,J. P.: MultiphaseandVolumetrict?quilibria

of Methane-n-HexancBinary System at Temperatures Between
- 110arrd15tYC,J. Chem. Erg. Data (Jan. 1%2)7, No. 1, 3-g.
14. Reamer, H, H., Sage, B. H., and Lacey, W. N.: PhaseEquilibria
in HydrocarbonSystems-Volumetric
andFlues Behaviorof the
Methane-rr-Heptane
System,Ind. i%. Chem. (1956)1, No. 1,
29-42.
ls< Kohn, J. P. and Bradish, W. F.: Mukiphascand Volumetric
Equilibria of the Methane-n-Octane System at Temperatures
Between -110 and 150C,J. Chem. Eng. Data (Jan. 1964)9,
No. 1,5-8.
16. Shipr&, L, M. and Kohn, P. J.: Heterogeneous
naw Md
Vohunetsic Equilibrium in the Methane-n-NotrarreSystem,
J. Chem. Eng. Data (April 1966)11,No. 2, 176-180.
17. Rrri, S. and Koho, J. P.: SoIid-Liquid-Va~r Equilibriumin the
Mctharr+n-l?ieosaneand Ethane-n-Eicosane Binary Systems,
J. Chem. Eng. Data (1970) 1S, No. 3,372-374.
alJi18. Sage, B. H., Webster, D. C., and Lacey,w. N.: phase
libna irr Hydrocarbon Systems,* lad. Eng. Chem. (1936) 28,
1045-1047.
19. Schoch, E. P., Hoffmarm, A. E., iksperick, A. S., Llghtfoot.

J. H,, and Maytlcld, F. D.: Solubdityof Methanein Benzene,

(June 1940)32,788-791.
20. Reanrer,H. H., Sage, B. H., and Lawy, W. N.: phase l%uilibria
in HydrocarbonSystems, Volumetric and Phase Behavior of the
Methane-CyclohexaneSystcm, Ind. Eng. Chem. (1958) 3, No.
2,240-245.
21. PostOn,R. S. and McKetta, J. J.: VapwLiquid Equilibrhmrin
the n-Hcxan~Nitrogen System,J. Chem. Eng. Data (July 1%6)
11, No. 3,364-365.
22. Azamoosh, A. and McKetta, J. J.: Nitrogen-n-DecaneSystem
1%3)8,
bt the Two-Phase Region. J. Chem. Eng. Data (Oct.
No. 4, 494-4%.
23. Roland, C. H.: Vapx-Liquid Equilibria for Natural Gas-Crrrde
Oil Mixtures,fnd. Eng. Chem. (Oct. 1945)37,N0. 10,930-936.
24, Standing, M. B. and Katz, D. L.: Vap-Liquid Equilibria of
Natuml Gas-Crude Oil Systems, Trans., AIME (1942) 155,
232-242.
25, Katz, D. L. and Hachmuth, D. H.: Vapnrimtion
r... __. .. . . %uilibrium
.
CQnstantsin a Crude Oil-NaturalGas Systcm, Ind. Eng. Chem.
(Sept. 1937)29,1072-1077.
Wapor-Iiquid Equilibriaof NaturalGas-Ctude
26. Standing, M. B...
011 Systems, PhD disscrtatiors, U. of Michigan.Ann&&
lnd. Eng. Chem.

(1941),
~bWmuwlptm@MWinSw~Vof
Petoleum Eng@eeraofliceSa@.15,1977.
Papar a~ad
tor @lcatii
Jan. 26, 197S. Reviead mmwwrlpt rernived July 19.
1978. PaLW (SPE ST21)first paeenwd a! UtaSPE.AIME 52rrdAnnualFall Taclm!cal
CcoferancaeumlExhbkbn. heldIn Denver,Ckt S-12, 1977.
ThmpqsarW ba hchdad inthe 1978 Transactions
WIluma.

165S